Structural and Solution Analysis for the Ordered Weber Problem under Spatial Uncertainty

TL;DR

This paper introduces a comprehensive analytical framework for single-facility location problems under spatial demand uncertainty, incorporating general probability measures and ordered weighted averaging, with theoretical properties and approximation schemes.

Contribution

It develops a unified stochastic ordered Weber model with fundamental properties, convergence analysis, and explicit error bounds, connecting convex analysis, stochastic programming, and ordered optimization.

Findings

Established convexity, continuity, and existence of solutions for the model.

Proved convergence and finite-sample error bounds for the approximation scheme.

Derived explicit error expressions for spherically symmetric distributions.

Abstract

We propose a general analytical framework for single-facility continuous location problems under spatial demand uncertainty. In contrast to classical formulations based on discrete or regionally aggregated demands, the proposed model represents uncertainty through general probability measures on $\R^d$, thereby encompassing finite, bounded, and unbounded support distributions within a unified formulation. The objective aggregates expected distances by means of an ordered weighted averaging operator, providing a flexible mathematical structure that includes the classical Weber problem and its ordered extensions as special cases. We establish fundamental properties of this stochastic ordered Weber model, including convexity, continuity, and existence of optimal solutions, and we derive quantitative bounds on the proximity between stochastic minimizers and the convex hulls of demand supports. Building upon these results, we develop and analyze an adaptive sample average approximation scheme, proving its convergence and deriving finite-sample error estimates under mild regularity conditions. For spherically symmetric distributions, we further obtain explicit analytical expressions for the approximation error. Together, these results provide a rigorous mathematical foundation for a broad class of stochastic ordered location models and highlight new theoretical connections between convex analysis, stochastic programming, and ordered optimization.